Studies on the relation between the pineal and human puberty have yielded conflicting results, partly due to methodologic issues, such as failure to consider the characteristic secretory rhythm of melatonin, or exclusion of normal subjects. Melatonin nocturnal rhythm was assessed in 62 normal subjects (ages 5-17 years), in whom hourly blood samples were collected by constant withdrawal from 18.00 to 08.00 h. Melatonin peak and mean integrated nocturnal concentrations were highly correlated (r = 0.90, p less than 0.05) and data are shown for melatonin peak. Tanner criteria were applied for pubertal stages. Mean melatonin peak was 153.6 +/- 72.6 pg/ml for stage 1, 141.0 +/- 26.2 pg/ml for stage 2, and 116.6 +/- 43.6 pg/ml for stages 3-5. There was a significant linear trend for decreasing melatonin peak with puberty. The subjects were also grouped into discrete age groups (5-6.99 years through 15-17 years); a significant linear trend was present for decreasing melatonin peak with age. The correlation between melatonin peak and age for all cases was r = -0.29, p less than 0.05. Analysis of covariance for melatonin peak among the pubertal groups showed no relation of melatonin and puberty when age was covaried. The time of peak was not significantly different among the pubertal groups or the age groups, despite a significant linear trend for later bedtime with age. The duration of the nocturnal surge was not significantly different among the pubertal groups. The data suggest that an overlapping effect of age may mask an interaction between the pineal and normal human puberty.